Both GR and QFT are insanely accurate models, but they are just models.
The N-body problem is undecidable, and Gödel, Turing, Church and other s proved that is the best we can do.
https://philsci-archive.pitt.edu/13175/1/parker2003.pdf
Western reductionism or Laplacian determinism is a good framework for practical, computable models. QFT actually is actually one of the counterexamples to Western reductionism.
But models are reductive and scientific models are just models. Don't confuse the map for the territory.
All models are wrong, some are useful; is another way of saying the same thing.
Superposition being inseparable is a large part on why the many words concept is popular with some people. It is about regaining a form of determinism.
Looking at gravity, we can compare what General Relativity and Quantum Mechanics say about the center of a black hole. Remember, both GR and QM have both been able to accurately model the way the world works at every scale we have been able to measure and test them. But they are incompatible with each other in certain points, such as the singularity in the black hole. GR says the center of the black hole is an infinitely dense point. QM says this can't be true because everything is made up of waves in a field, which requires things be spread out over some amount of an area. These can't both be true, yet GR and QM have both stood up to every single test and observation we can throw at them. Every prediction they make that we can verify has been verified and lines up with the theories. And this is not the only place they disagree, of course, but it is one example.
And that's really, from my understanding, the more fundamental answer to the question asked in the reddit post. It's not that unifying the two requires a gauge boson like the graviton, though that it is one possible outcome when quantizing gravity, but that we have two very useful and very tested models of how things work that are incompatible with each other in certain ways. Maybe gravitons exist, though it currently seems impossible for us to reach the point where we can detect them - Dyson calculated that using an Earth sized detector we'd be able to detect about one graviton from the sun per billion years, if they exist - and maybe they don't.
As for predicted things not expecting to exist in reality, this is really just par for the course for models. It's not like tensors are some real physical thing either, for example.
Approximately: consider gravity as small deviations from inertiality, so it's approximately linear, like the other forces we know and love.
A. When you incessantly hunt for every boson you can find anywhere, you end up with more of them.